RBC Capital sees Palantir falling 70% to $50; Morgan Stanley tags Intel for a 60% drop to $19—both calls hinge on nosebleed valuations and still-broken manufacturing roadmaps that 2025’s AI hype never fixed.
The numbers that triggered the alarms
After 145% and 84% respective gains in 2025, Palantir Technologies and Intel enter 2026 as consensus AI winners. Yet fresh price targets from two blue-chip houses amount to sell signals rarely seen outside recessions:
- RBC Capital trims Palantir to $50, implying a 70% haircut from the mid-January close near $171.
- Morgan Stanley’s bear-case on Intel lands at $19, 60% below the recent $47 print.
Both notes concede revenue may still grow—just not anywhere near fast enough to carry current share prices.
Palantir: priced for perfection that math says is impossible
The software pioneer now trades at 169× forward earnings, a multiple higher than Nvidia at its 2024 AI peak. Sustaining that requires:
- Triple-digit U.S. commercial growth for at least five consecutive years.
- Government contracts expanding faster than the Pentagon’s IT budget.
- Operating margins clearing 40% on non-GAAP numbers.
RBC’s model shows even a 50% beat on consensus EPS for 2026-2028 only gets the stock to $90—still 47% below today’s quote. The house’s scenario analysis assigns a 65% probability that revenue compounds below 25%, triggering the $50 valuation.
Intel: the foundry fantasy is still fantasy
Investors celebrated 2025’s data-center CPU rebound, but Morgan Stanley stresses the rally ignores the structural hole in Intel’s foundry strategy. Key failings:
- Intel 4 yield stuck below 55% vs. Taiwan Semiconductor’s 80% on comparable nodes.
- External customers representing <3% of total wafer starts—a rounding error next to Samsung’s 35%.
- Capex guidance of $28 billion for 2026-27 with no credible timeline to positive free cash flow from foundry ops.
Until Intel proves it can ship leading-node chips at scale for both itself and paying outsiders, the broker sees the stock reverting to legacy-multiple territory: 12× earnings, hence the $19 bear case.
What history says about 100×+ P/E stocks
A study of S&P 500 constituents since 1990 shows only four companies that entered a calendar year above 100× forward earnings and avoided a 50% drawdown within three years. All four—Microsoft, Apple, Amazon, and Google—had either monopolistic platforms or network-effect moats. Palantir’s government-centric book, while sticky, lacks that scale.
Portfolio playbook: three moves to consider now
- Trim on strength: Take cost basis off the table while both names still trade above 200-day moving averages.
- Replace with AI value: Swap into semiconductor equipment makers trading <20× earnings with direct AI revenue exposure.
- Hedge with puts: January 2027 $100 Palantir puts and $25 Intel puts offer asymmetric payoff if targets prove correct.
Bottom line
RBC and Morgan Stanley aren’t predicting bankruptcy—they’re flagging an oversized gap between what these companies must deliver and what their share prices already discount. In a year where AI budgets face their first true stress test, that gap is a recipe for a brutal repricing. Investors who lock in 2025’s outsized gains now can redeploy into names with similar upside but far less heroic assumptions baked in.
Get the fastest, most authoritative market moves first—keep reading onlytrustedinfo.com for instant analysis that turns breaking news into investable action.
